RBMRB : A package to download and visualize BMRB data in R
Kumaran Baskaran
November 08, 2016
RBMRB
BMRB collects, annotates, archives, and disseminates (worldwide in the public domain) the important spectral and quantitative data derived from NMR spectroscopic investigations of biological macromolecules and metabolites. The goal is to empower scientists in their analysis of the structure, dynamics, and chemistry of biological systems and to support further development of the field of biomolecular NMR spectroscopy
RBMRB is a library to fetch NMR chemical shift data directly from BMRB into R environment as a data frame in R. This facilitates access to BMRB data for statistical analysis and data visualization. It is using the BMRB-API to fetch the data from BMRB database.
Installation
RBMRB library has been developed and tested in R version 3.3.x. It requires the following R packages preinstalled
- httr to import data from BMRB web server(version 1.2.1 or later)
- data.table to format the imported data into a data frame in R (version 1.9.6 or later)
- rjson to deal with BMRB-API (version 0.2.15 or later)
- ggplot2 to simulate spectra (version 2.1.0 or later)
- plotly for interactive graphics in simulated spectra (version 4.5.2 or later)
Users should make sure that the above packages have beed installed correctly with the required versions, before proceeding to RMBRM insallation.
Here is the instruction to install those packages. Open your R and use the following command
installed.packages(c("httr","data.table","rjson","ggplot2","plotly"))Method 1
Once the necessary packages have been installed, proceed with RMRBM installation. The source file can be downloaded from GitHub
After downloading the source file, use the following command to install RBMRB library
install.packages("~/Downloads/RBMRB_2.0.tar.gz",repos=NULL,type="source")Note: provide the correct path to the downloaded file.
Method 2
If you have devtools library in your R, then you can install directly from GitHub.
library(devtools)
install_github("uwbmrb/RBMRB/RBMRB")Usage
RBMRB can be used in a similar way like any other library in R.
library(RBMRB)Data access
BMRB data can be imported in two ways
- Entry method Chemical shift data from single or multiple entries
- Atom method Chemical shift data from all entries for a given atom
Entry method
fetch_entry_chemical_shifts:
This function will fetch the ‘Atom_chem_shift’ loop from a NMR-STAR file for a given entry or a list of entries in CSV format. This function works on both macromolecules and metabolites data base. For metabilites entry ids should have right prefix (example ‘bmse000034’)
Examples:
df1<-fetch_entry_chemical_shifts(15060)
df2<-fetch_entry_chemical_shifts(c(17074,17076,17077))
df2<-fetch_entry_chemical_shifts(c('17074','17076','17077'))
df3<-fetch_entry_chemical_shifts(c('bmse000034','bmse000035','bmse000036'))These data frames have the following columns
colnames(df1)## [1] "ID" "Assembly_atom_ID"
## [3] "Entity_assembly_ID" "Entity_ID"
## [5] "Comp_index_ID" "Seq_ID"
## [7] "Comp_ID" "Atom_ID"
## [9] "Atom_type" "Atom_isotope_number"
## [11] "Val" "Val_err"
## [13] "Assign_fig_of_merit" "Ambiguity_code"
## [15] "Occupancy" "Resonance_ID"
## [17] "Auth_entity_assembly_ID" "Auth_asym_ID"
## [19] "Auth_seq_ID" "Auth_comp_ID"
## [21] "Auth_atom_ID" "Details"
## [23] "Entry_ID" "Assigned_chem_shift_list_ID"Sample data output
head(df1)## ID Assembly_atom_ID Entity_assembly_ID Entity_ID Comp_index_ID Seq_ID
## 1 1 . 1 1 20 20
## 2 2 . 1 1 20 20
## 3 3 . 1 1 20 20
## 4 4 . 1 1 20 20
## 5 5 . 1 1 21 21
## 6 6 . 1 1 21 21
## Comp_ID Atom_ID Atom_type Atom_isotope_number Val Val_err
## 1 LEU H H 1 8.149 NA
## 2 LEU CA C 13 56.016 NA
## 3 LEU CB C 13 42.180 NA
## 4 LEU N N 15 122.739 NA
## 5 VAL H H 1 8.048 NA
## 6 VAL CA C 13 63.412 NA
## Assign_fig_of_merit Ambiguity_code Occupancy Resonance_ID
## 1 . 1 . .
## 2 . 1 . .
## 3 . 1 . .
## 4 . 1 . .
## 5 . 1 . .
## 6 . 1 . .
## Auth_entity_assembly_ID Auth_asym_ID Auth_seq_ID Auth_comp_ID
## 1 . . 20 LEU
## 2 . . 20 LEU
## 3 . . 20 LEU
## 4 . . 20 LEU
## 5 . . 21 VAL
## 6 . . 21 VAL
## Auth_atom_ID Details Entry_ID Assigned_chem_shift_list_ID
## 1 HN . 15060 1
## 2 CA . 15060 1
## 3 CB . 15060 1
## 4 N . 15060 1
## 5 HN . 15060 1
## 6 CA . 15060 1Atom method
fetch_atom_chemical_shifts:
This function will fetch the chemical shift data from all the entries for a given atom. The atom name should be in NMR-STAR atom nomenclature.
Examples:
df4<-fetch_atom_chemical_shifts('CG2')
df5<-fetch_atom_chemical_shifts('C9')These data frames have the following columns
colnames(df4)## [1] "Entry_ID" "Entity_ID"
## [3] "Comp_index_ID" "Comp_ID"
## [5] "Atom_ID" "Atom_type"
## [7] "Val" "Val_err"
## [9] "Ambiguity_code" "Assigned_chem_shift_list_ID"Sample data output
head(df4)## Entry_ID Entity_ID Comp_index_ID Comp_ID Atom_ID Atom_type Val
## 1 10001 1 1 ILE CG2 C 15.700
## 2 10001 1 6 ILE CG2 C 17.900
## 3 10002 1 3 ILE CG2 C 17.516
## 4 10002 1 18 VAL CG2 C 22.278
## 5 10002 1 19 THR CG2 C 21.957
## 6 10002 1 26 THR CG2 C 21.779
## Val_err Ambiguity_code Assigned_chem_shift_list_ID
## 1 0.4 1 1
## 2 0.3 1 1
## 3 0.4 1 1
## 4 0.4 1 1
## 5 0.4 1 1
## 6 0.4 1 1Data manipulation
There are few data manipulation functions are availbale to facilitate plotting.
convert_cs_to_n15hsqc:
This function will reformat the chemical shift data frame into a data frame which is easy to plot the N15-HSQC spectrum from the data.
Examples
n15hsqc1<-convert_cs_to_n15hsqc(df1)
n15hsqc2<-convert_cs_to_n15hsqc(df2)The output data frame will look like
head(n15hsqc1)## Entry_ID Comp_index_ID Entity_ID Assigned_chem_shift_list_ID Comp_ID_H
## 1 15060 101 1 1 ASP
## 2 15060 102 1 1 ASP
## 3 15060 103 1 1 SER
## 4 15060 104 1 1 ASP
## 5 15060 105 1 1 GLU
## 6 15060 106 1 1 GLU
## Comp_ID_N H N
## 1 ASP 8.269 120.647
## 2 ASP 8.376 120.080
## 3 SER 8.239 115.602
## 4 ASP 8.409 122.658
## 5 GLU 8.269 120.647
## 6 GLU 8.391 122.119This data frame is easy to plot using any plotting library
library(ggplot2)
plt1<-ggplot(n15hsqc1)+geom_point(aes(x=H,y=N))
plt1
plt2<-ggplot(n15hsqc2)+geom_point(aes(x=H,y=N,color=Entry_ID))
plt2
plt3<-ggplot(n15hsqc2)+geom_point(aes(x=H,y=N,color=Entry_ID))+geom_line(aes(x=H,y=N,group=Comp_index_ID))
plt3
convert_cs_to_c13hsqc:
This function will reformat the chemical shift data frame into a data frame which is easy to plot the C13-HSQC spectrum from the data.
Examples
c13hsqc1<-convert_cs_to_c13hsqc(df1)
c13hsqc2<-convert_cs_to_c13hsqc(df2)The output data frame will look like
head(c13hsqc1)## Entry_ID Comp_index_ID Entity_ID Assigned_chem_shift_list_ID Comp_ID_C
## 1 15060 101 1 1 ASP
## 2 15060 102 1 1 ASP
## 3 15060 103 1 1 SER
## 4 15060 104 1 1 ASP
## 5 15060 105 1 1 GLU
## 6 15060 106 1 1 GLU
## Comp_ID_H Atom_ID_C Atom_ID_H C H
## 1 ASP CA HA 54.487 4.630
## 2 ASP CA HA 54.572 4.609
## 3 SER CA HA 58.470 4.420
## 4 ASP CA HA 54.567 4.640
## 5 GLU CA HA 56.521 4.271
## 6 GLU CA HA 56.400 4.300and the user may generate a spectrum using the following script
library(ggplot2)
plt1<-ggplot(c13hsqc1)+geom_point(aes(x=H,y=C))
plt1
plt2<-ggplot(c13hsqc2)+geom_point(aes(x=H,y=C,color=Entry_ID))
plt2
convert_cs_to_tocsy:
This function will reformat the chemical shift data frame into a data frame which is easy to plot the TOCSY spectrum from the data. Note : Since both dimensions have protein chemical shifts, the columns are named as Val.x and Val.y
Examples
tocsy1<-convert_cs_to_tocsy(df1)
tocsy2<-convert_cs_to_tocsy(df2)after conversion the data will look like
head(tocsy1)## Entry_ID Entity_ID Comp_index_ID Assigned_chem_shift_list_ID ID.x
## 1 15060 1 100 1 915
## 2 15060 1 100 1 915
## 3 15060 1 100 1 916
## 4 15060 1 100 1 916
## 5 15060 1 101 1 919
## 6 15060 1 101 1 919
## Assembly_atom_ID.x Entity_assembly_ID.x Seq_ID.x Comp_ID.x Atom_ID.x
## 1 . 1 100 GLY HA2
## 2 . 1 100 GLY HA2
## 3 . 1 100 GLY HA3
## 4 . 1 100 GLY HA3
## 5 . 1 101 ASP H
## 6 . 1 101 ASP H
## Atom_type.x Atom_isotope_number.x Val.x Val_err.x Assign_fig_of_merit.x
## 1 H 1 3.960 NA .
## 2 H 1 3.960 NA .
## 3 H 1 4.000 NA .
## 4 H 1 4.000 NA .
## 5 H 1 8.269 NA .
## 6 H 1 8.269 NA .
## Ambiguity_code.x Occupancy.x Resonance_ID.x Auth_entity_assembly_ID.x
## 1 2 . . .
## 2 2 . . .
## 3 2 . . .
## 4 2 . . .
## 5 1 . . .
## 6 1 . . .
## Auth_asym_ID.x Auth_seq_ID.x Auth_comp_ID.x Auth_atom_ID.x Details.x
## 1 . 100 GLY HA1 .
## 2 . 100 GLY HA1 .
## 3 . 100 GLY HA2 .
## 4 . 100 GLY HA2 .
## 5 . 101 ASP HN .
## 6 . 101 ASP HN .
## ID.y Assembly_atom_ID.y Entity_assembly_ID.y Seq_ID.y Comp_ID.y
## 1 915 . 1 100 GLY
## 2 916 . 1 100 GLY
## 3 915 . 1 100 GLY
## 4 916 . 1 100 GLY
## 5 919 . 1 101 ASP
## 6 920 . 1 101 ASP
## Atom_ID.y Atom_type.y Atom_isotope_number.y Val.y Val_err.y
## 1 HA2 H 1 3.960 NA
## 2 HA3 H 1 4.000 NA
## 3 HA2 H 1 3.960 NA
## 4 HA3 H 1 4.000 NA
## 5 H H 1 8.269 NA
## 6 HA H 1 4.630 NA
## Assign_fig_of_merit.y Ambiguity_code.y Occupancy.y Resonance_ID.y
## 1 . 2 . .
## 2 . 2 . .
## 3 . 2 . .
## 4 . 2 . .
## 5 . 1 . .
## 6 . 1 . .
## Auth_entity_assembly_ID.y Auth_asym_ID.y Auth_seq_ID.y Auth_comp_ID.y
## 1 . . 100 GLY
## 2 . . 100 GLY
## 3 . . 100 GLY
## 4 . . 100 GLY
## 5 . . 101 ASP
## 6 . . 101 ASP
## Auth_atom_ID.y Details.y
## 1 HA1 .
## 2 HA2 .
## 3 HA1 .
## 4 HA2 .
## 5 HN .
## 6 HA .Plotting TOCSY spectrum
library(ggplot2)
plt1<-ggplot(tocsy1)+geom_point(aes(x=Val.x,y=Val.y))
plt1
plt2<-ggplot(tocsy2)+geom_point(aes(x=Val.x,y=Val.y,color=Entry_ID))
plt2
filter_residue:
This function will filter the data frame and remove all non standard amino acids. The data frame should contain the amino acid information in the Comp_ID column. ####Examples
df6<-fetch_atom_chemical_shifts('CG2')
df7<-filter_residue(df6)Data visualization
RBMRB library contains few functions to generate interactive visualization of BMRB data with out any data manipulation. The interactive visualizations use plotly library. If user has problem with plotly, then this feature may be diabled by providing an argument ‘interactive=FALSE’ for these functions. These interactive plots can be zoomed in and out using a mouse and will show tooltip information when you mouse over. These visualizations can be exported as a stand alone html file
HSQC_15N
This function will simulae N15-HSQC spectrum for a given entry or list of entries.
Examples
These interactive visualization can be exported as single stand alone html file
spec1<-HSQC_15N(15060)
spec1spec2<-HSQC_15N(c(17074,17076,17077),type='line')
spec2spec3<-HSQC_15N(c(17074,17076,17077),type='line',interactive = F)
spec3
HSQC_13C
This function will simulae C13-HSQC spectrum for a given entry or list of entries.
Examples
These interactive visualization can be exported as single stand alone html file
spec1<-HSQC_13C(15060)
spec1spec2<-HSQC_13C(c(17074,17076,17077))
spec2spec3<-HSQC_13C(c(17074,17076,17077),interactive = F)
spec3
TOCSY
This function will simulae TOCSY spectrum for a given entry or list of entries.
Examples
These interactive visualization can be exported as single stand alone html file
spec1<-TOCSY(15060)
spec1spec2<-TOCSY(c(17074,17076,17077))
spec2spec3<-TOCSY(c(17074,17076,17077),interactive = F)
spec3
chemical_shift_corr
This function will plot the distribution of checmial shift correlation between any two atoms from the 20 standard amino acids. The distribution of a particular residue may turn on and off by clicking the residue name in the legand.
corr_plot1<-chemical_shift_corr('CB','N')
corr_plot1